JP2008139085A - Pattern dimension measuring method, and scanning transmission charged particle microscope - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a pattern dimension measuring method capable of measuring a cross-sectional shape using a scanning charged particle microscope such as a STEM (Scanning Transmission Electron Microscope), as to a plurality of cross sections, by one time of sample preparation, and a system therefor. <P>SOLUTION: A focal position of a focusing charged particle ray is sequentially changed along a z-direction, following the desired cross section of a measuring objective pattern, a transmission electron image or a scattering electron image of the measuring objective pattern is acquired in each focal position, the transmission electron image or the scattering electron image acquired in each focal position is processed to calculate an edge position of the measuring objective pattern within the electron image in each focal position, and the three-dimensional cross-sectional shape of the measuring objective pattern is measured based on a combination of the calculated edge position of the measuring objective pattern within the electron image in each focal position, and each focal position, using the scanning charged particle microscope. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a pattern dimension measuring method and a scanning transmission charged particle microscope for measuring a dimension of a processed shape of a circuit pattern formed on a wafer in a semiconductor manufacturing process using a transmission or scattering charged particle microscope image of the circuit pattern. And the system.

  In the manufacturing process of a semiconductor wafer, the dimension and shape of a pattern formed in multiple layers on the wafer are one of the important factors that determine the performance of the semiconductor device. In particular, wiring patterns including transistor gate wirings are strongly related to the wiring width and shape and device operation characteristics, and it is very important to construct a manufacturing process that realizes a desired wiring shape.

  As a length measurement tool for measuring the line width of fine wiring of the order of several tens of nanometers, a scanning electron microscope (line length measuring SEM) for measuring the width of the wiring capable of imaging the wiring at a magnification of 100,000 to 200,000 times. (Scanning Electron Microscope) or CD (Critical dimension) SEM) has been conventionally used. However, since it is difficult to confirm the shape of the pattern cross-section with a length measurement SEM, a patterned wafer is cleaved or a focused ion beam (FIB) processing apparatus at the initial stage of the start-up of the manufacturing process. In general, the pattern shape is processed by using SEM, Transmission Electron Microscope (TEM) or Scanning Transmission Electron Microscope (STEM) to confirm the cross-sectional shape. Has been done. The length measurement SEM is described in, for example, Japanese Patent Laid-Open No. 2005-265424 (Patent Document 2).

  Conventionally, this cross-sectional shape evaluation has been confirmed by the observer by the appearance of the shape, but with the recent miniaturization of semiconductor devices, the demand for high-precision dimension measurement is also increasing for the cross-sectional shape. . However, there is a problem that the cross-sectional measurement requires a lot of time of several tens of minutes to several hours for the sample preparation. A normal semiconductor circuit pattern has variations in pattern edge position and dimension called line edge roughness (LER) and line width roughness (LWR). In cross-sectional observation, since it is possible to observe only one place that is usually processed, it is necessary to observe pattern dimensions at many points in order to evaluate the state of the entire process for patterns with large roughness. There is. There is an atomic force microscope (AFM) as another device that measures the shape of the pattern cross section, but as shown in Non-Patent Document 1, it can be measured well depending on the thickness and shape of the probe. Sometimes not. In particular, it is difficult to measure a shape in which the bottom portion of the wiring is thinner than the upper portion, which is called reverse taper or notch, and it cannot be said that it can sufficiently cope with future pattern miniaturization.

US6, 548, 810 B2 Specification JP 2005-265424 A B. Banke, C. Archie, M. Sendelbach et al, “Reducing measurement uncertainty drives the use of multiple technologies for supporting metrology” Proc. SPIE 2375, pp. 133-150 (2006) A. Yamaguchi, H. Fukuda, H. Kawada and T. Iizumi, “Impact of Long-Period Line-Edge Roughness (LER) on Accuracy in CD Measurement” Proc. SPIE 5752, pp.1362-1370 (2005) H. Okushima, T. Onozuka et al, “ArF photo resist pattern sample preparation method using FIB without protective coating” Proc. SPIE 6152-165 (2006) Japan Electron Microscope Society Kanto Branch, Scanning Electron Microscope, pp83-86 K. van Benthem, S.N.Rashkeev and S.J.Pennycook, “Atomic and Electronic Structure Investigations of HfO2 / SiO2 / Si Gate Stacks Using Aberration-Corrected STEM” Characterization and Metrology for ULSI Technology 2005, pp79-84 J. Notte and B. Ward, “Sample interaction and contrast mechanisms of the helium ion microscope” SCANNING, pp.63-64, 2006

  As shown in the background art, when cross-sectional measurement is performed by SEM, TEM, or STEM, it takes a lot of time to create a sample, and it is difficult to measure many cross-sections.

  An object of the present invention is to provide a pattern dimension measuring method capable of performing cross-sectional shape measurement using a scanning transmission charged particle microscope such as a STEM (Scanning Transmission Electron Microscope) on a plurality of cross-sections by creating a sample once. An object of the present invention is to provide a scanning transmission charged particle microscope such as STEM and a system thereof. In other words, the object of the present invention is to create a sample that requires a lot of time in one time, and to substantially reduce the substantial throughput of cross-sectional shape measurement using a scanning charged particle microscope such as STEM. An object of the present invention is to provide a pattern dimension measuring method, a scanning transmission charged particle microscope such as STEM, and a system thereof.

  In addition, another object of the present invention is to enable many cross-sectional shape measurements in the same required time as a normal cross-sectional shape measurement by using a scanning transmission charged particle microscope such as STEM, and for a rough pattern. It is possible to easily estimate an average cross-sectional dimension, and to provide a pattern dimension measuring method and system useful for process evaluation and comparison with other measuring apparatuses (for example, length measuring SEM). .

  In order to achieve the above object, the present invention provides a pattern measurement method for measuring a three-dimensional cross-sectional shape of a measurement target pattern having a relatively large length of 2 μm or more in the z direction using a scanning charged particle beam microscope. The focus position of the focused charged particle beam is sequentially changed with respect to the z direction of the measurement target pattern by matching the scanning irradiation position of the focused charged particle beam with the desired xy cross-section measurement position of the measurement target pattern, An electronic image acquisition process for acquiring a transmission electron image or a scattered electron image of the measurement target pattern at each focal position, and at each focal position that matches the desired xy cross-sectional measurement position in the electronic image acquisition process The obtained transmission electron image or scattered electron image is processed to calculate the edge position of the measurement target pattern in the electronic image at each focal position. And based on the combination of the edge position of the measurement target pattern in the electronic image at each focal position calculated in the edge position calculation process and the respective focal position from which the electronic image was acquired in the electronic image acquisition process, A pattern measuring method comprising: a cross-sectional shape measuring process for measuring a three-dimensional cross-sectional shape of the measurement target pattern.

  In the invention, it is preferable that the measurement target pattern is formed as a line pattern in the z direction. According to the present invention, the measurement target pattern is formed as a line pattern having a length of about 2 μm or more in the z direction, and in the cross-sectional shape measurement process, a three-dimensional cross-sectional shape measurement of the measurement target pattern is performed in a predetermined z cross section. The average dimension, line width roughness or line edge roughness of the line pattern is measured.

  Further, the present invention is characterized in that a solid such as a protective film is not attached around the measurement target pattern. In the electronic image acquisition process according to the present invention, an acceleration voltage of the charged particle beam in the scanning charged particle beam microscope is 200 keV or more. In the electronic image acquisition process, the present invention further includes a beam opening angle determination process for determining a beam opening angle of the charged particle beam in the scanning charged particle beam microscope based on a dimensional change amount of the measurement target pattern. It is characterized by including.

  Further, in the edge position calculation process, the present invention is configured such that, in the signal waveform obtained from the electronic image at each focal position, the coordinate having the maximum inclination is set as the edge coordinate of the measurement target pattern. Features. Further, according to the present invention, in the edge position calculation process, in the signal waveform obtained from the electronic image at each focal position, the signal waveform matches a threshold value set according to the image acquisition condition. Is the edge coordinates of the measurement target pattern.

  Further, the present invention provides a relationship between the calculation result of the edge position of the measurement target pattern in the electronic image at each focal position calculated in the edge position calculation process and the respective focal position in the cross-sectional shape measurement process. Is included in a linear expression, and an inclination correction process is included in which the calculation result of the edge position of the measurement target pattern at each focus position is corrected by the inclination of the approximate expression.

  Further, in the cross-sectional shape measurement process, the present invention provides a difference from an average value of the calculation result of the edge position of the measurement target pattern in the electronic image at each focal position calculated in the edge position calculation process in advance. When it is larger than the specified allowable value, an exclusion process for excluding the measurement result of the three-dimensional cross-sectional shape of the measurement target pattern is included.

  The present invention also relates to a pattern measuring method for measuring a three-dimensional cross-sectional shape of a measurement target pattern using a scanning charged particle beam microscope, wherein the scanning irradiation position of the converged charged particle beam is determined as desired for the measurement target pattern. The focus position of the converged charged particle beam is sequentially changed in the z direction with respect to the measurement target pattern in accordance with the xy cross-section measurement position of the first transmission electron image of the measurement target pattern at each focus position. Alternatively, a first electronic image acquisition process for acquiring a scattered electron image, and a first transmission electron image acquired at each focal position in accordance with the desired xy cross-sectional measurement position in the first electronic image acquisition process Alternatively, a first edge position calculation process of processing a scattered electron image to calculate an edge position of a measurement target pattern in the electronic image at each focal position, and the measurement target pattern And the focus irradiation position of the focused charged particle beam is set to z with respect to the inverted measurement target pattern by matching the scanning irradiation position of the focused charged particle beam with the desired xy sectional measurement position of the inverted measurement target pattern. A second electronic image acquisition process in which the second transmission electron image or the scattered electron image of the inverted measurement target pattern at each focal position is acquired by sequentially changing the direction, and in the second electronic image acquisition process The second transmission electron image or the scattered electron image acquired at each focal position matched with the desired xy cross-sectional measurement position is processed, and the measurement target pattern inverted in the electronic image at the respective focal position is processed. Measurement target pattern in the electronic image at each of the focal positions calculated in the second edge position calculation process for calculating the edge position of the image and the first edge position calculation process A combination of an edge position and each focal position from which the electronic image was acquired in the first electronic image acquisition process, and an inverted measurement in the electronic image at each focal position calculated in the second edge position calculation process A cross-sectional shape measurement process for measuring a three-dimensional cross-sectional shape of the measurement target pattern based on a combination of an edge position of the target pattern and each focus position from which the electronic image is acquired in the second electronic image acquisition process. This is a pattern measurement method characterized by this.

  According to the present invention, in the cross-sectional shape measurement process, the calculation result of the edge position of the measurement target pattern calculated in the first edge position calculation process and the inverted measurement calculated in the second edge position calculation process. The three-dimensional cross-sectional shape measurement of the measurement target pattern is performed by a weighted average obtained by applying a weight corresponding to the depth from the surface of the measurement target pattern to the calculation result of the edge position of the target pattern.

  In the scanning charged particle beam microscope, the present invention provides a longitudinal dimension of a measurement target pattern for acquiring a transmission electron image or a scattered electron image of the measurement target pattern in an electronic image acquisition unit and a focal position to be sequentially changed. The image processing apparatus includes a graphical user interface for inputting and setting an image acquisition condition including an interval and an image processing condition for calculating an edge position of the measurement target pattern in the image processing unit.

  As described above, according to the present invention, it is possible to realize pattern measurement capable of measuring many cross-sectional shapes and dimensions with one sample preparation.

  In addition, according to the present invention, many cross-sectional shape measurements can be performed in a relatively short time. Thereby, it is possible to evaluate the manufacturing variation of the cross-sectional shape, which has been difficult until now, for the semiconductor circuit pattern to be measured. Especially for shape changes such as reverse taper and notch that are difficult to grasp with the non-destructive measurement method, high-precision measurement is possible by using a STEM image with good resolution.

  Further, according to the present invention, as a result of enabling such evaluation, the state of the semiconductor process, particularly the state of the process start-up period can be quantitatively evaluated in a short period of time, and the start-up period can be shortened. It becomes possible.

  In addition, according to the present invention, statistical evaluation in consideration of variations in cross-sectional shape and dimensions can be performed, and therefore, it can be applied to calibration of various measuring devices used in semiconductor lines. .

  DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of a pattern dimension measuring method and system for measuring a dimension of a processed shape of a circuit pattern formed on a wafer in a semiconductor manufacturing process according to the present invention using an electron microscope image of the circuit pattern are described with reference to the drawings. I will explain.

[First Embodiment]
As a first embodiment, a basic procedure and system configuration of a pattern measurement method according to the present invention will be described with reference to FIGS.

  FIG. 1 shows the basic procedure. First, a sample is prepared by removing the periphery using, for example, a focused ion beam (FIB) processing apparatus so that the cross-section can be observed (S01). In the present invention, dimension measurement is performed using a sample having a larger (thick) dimension in the longitudinal direction of a line whose dimension changes than that used in normal cross-sectional observation. FIG. 2 shows an example of cutting out a sample for cross-sectional observation. When the sample is cut out for measuring the cross-sectional shape, the length L of the pattern to be measured shown in FIG. 2 is recorded.

  FIG. 2A shows a wiring circuit pattern to be measured as viewed from above. The white portion is the line 11 and the gray portion is the base 12. In FIG. 2, the line 11 is shown as a straight line for simplicity, but the line width of the semiconductor line pattern 11a on the actual base 12 varies depending on roughness as shown in FIG. Such line pattern roughness is often caused by an exposure process in a semiconductor manufacturing process. This roughness is one of the main causes of performance variation of semiconductor devices, and it is very important to measure this. The present invention provides means for measuring the cross-sectional shape at the same time as observing the roughness using a scanning transmission charged particle microscope such as STEM.

  FIG. 2B shows an example of a sample obtained by processing a region indicated by a rectangle with, for example, FIB. Usually, the cross-sectional shape is measured on the surface shown by hatching in FIG. 2 (b). In the measurement method using the STEM of the present invention, an electron beam is irradiated perpendicularly to this surface, Perform dimension measurements on multiple parallel sections. Usually, when measuring cross-sections using STEM or TEM, the pattern length (the part indicated by the arrow in FIG. 2B) should be relatively thin, from several tens of nanometers to several hundred nanometers at most. Is. However, in the present invention, a relatively long sample of about several micrometers (about 2 μm or more) including a plurality of cross sections to be measured is created, and the dimensions and roughness in the created sample are measured. As described in Non-Patent Document 2, when considering the influence on a semiconductor device, it is recommended that roughness measurement be performed in a range of 2 micrometers or more.

  The present invention discloses a method that enables roughness measurement at 2 micrometers.

  For the sake of simplicity, FIG. 2 shows a diagram in which only four lines are included in one sample. Of course, more lines may be cut out at a time. Depending on the pattern density, for example, a region including several tens of lines is cut out at a time. In this case, there is an advantage that a large number of cross sections can be measured by changing the image acquisition position by creating a sample once. Of course, a protective film may be generated in order to reduce damage in the FIB processing. As will be described later, the measurement method of the present invention provides a more accurate result when there is no protective film around the pattern to be measured, but can be applied even with a protective film. When using a transmission electron image, it is desirable to use a film having an atomic number as small as possible as compared with the target pattern so that electrons can be easily transmitted. Note that Non-Patent Document 3 discloses a method for creating a sample that does not use a protective film by FIB processing. The following description will be given in the case where there is no protective film, but the cross-sectional dimension can be measured by the same procedure even if the protective film is present.

  Next, for example, in a STEM (Scanning Transmission Electron Microscope) shown in FIG. 4, the prepared sample is set in a measurement holder (not shown), and axis adjustment, sample inclination adjustment, magnification adjustment, etc. are performed. The STEM is adjusted so that the cross-section can be measured (FIG. 1, S02).

  FIG. 4 shows an example of the structure of the STEM used in the present invention. In the STEM, electrons emitted from the electron source 301 are accelerated, and the beam 308 is converged finely through a plurality of stages of the electron lens 302 and the objective lens 304. An image signal at each point is detected by causing the electron beam 308 to scan in a rectangle using the scanning coil 303 while irradiating the sample (sample) 305. In order to obtain transmitted electrons, an acceleration voltage of electrons irradiated to the sample set by the acceleration voltage setting unit 300b is several hundred keV that is 200 keV or more. The electrons irradiated on the sample 305 are transmitted and scattered in the sample and detected by a lower detector. Although the form of the detector varies depending on the purpose, FIG. 5 shows a case where there are a general bright field detector 307, dark field detector 306, and secondary electron detector 310. The dark field detector 306 uses an annular type detector as shown in FIG. 5, and detects only electrons having a larger scattering angle. Since the dark field image well represents the difference in atomic number, it is also called Z contrast. The bright field detector 307 is disposed below the dark field detector 306, and the bright field detector 307 is not scattered so much and detects mainly the electrons transmitted below the sample 305. Also, as shown in FIG. 4, there is an apparatus in which a secondary electron detector 310 is provided above the sample so that secondary electrons can be detected. Electrons detected by these detections are converted into digital signals by an A / D converter 309, transferred to an image processing unit 314 having an image memory and a CPU for image processing, and recorded as image data in a storage device 311. The recorded image data is processed by the image processing unit 314. Alternatively, external processing can be performed through the external interface 312. Device scanning and result display are performed through the display device 313. The STEM electron optical system is controlled by the controller 300. The control device 300 is observed by an electron beam opening angle adjusting unit 300a that adjusts a focal depth mode (electron beam opening angle α) corresponding to the resolution in the z direction, an acceleration voltage setting unit 300b, and a secondary electron detector 310. A focus adjustment position for controlling the objective lens 304 or the like to a focus adjustment position according to the measurement interval p and the sample length L as shown in FIG. A control unit 300c and a beam scanning irradiation position control unit 300d that controls the deflector 303 and the like are provided. The beam scanning irradiation position control unit 300d includes a part for controlling the magnification of the acquired image.

  Patent Document 1 discloses an example of a confocal microscope as a system similar to the present invention. In Patent Document 1, an electron source, a sample, and a detector are arranged at optically conjugate positions, and an image inside the object is formed and observed.

  In the present invention, since the edge position of the pattern is detected using the difference in the degree of transmission and scattering depending on the pattern, it is not necessary to form an electron beam image on the detector.

  Next, an image for cross-sectional measurement at a desired cross-sectional measurement position is acquired using the system shown in FIG. In the present invention, an embodiment is shown in which measurement is performed mainly using a transmission electron image (bright field image) at a desired cross-sectional measurement position. Images obtained by changing the focus in the longitudinal direction (z direction shown in FIG. 2B) of the line pattern to be cross-sectional measured are obtained, and the edge position of the pattern is calculated in each image. Although details will be described later, in this method, it is possible to estimate the pattern cross-sectional shape and its dimensions at the focused position. In order to acquire an image with the focus changed, the focus position is adjusted (S03 in FIG. 1). Here, an embodiment of a STEM capable of detecting a secondary electron image is shown. Here, first, a focus condition in which the secondary electron image is best seen is set using the secondary electron image. Since the secondary electron image contains a lot of information on the surface of the sample, the electron beam is focused on the outermost surface of the sample under the conditions determined using the secondary electron image. If the secondary electron image cannot be detected, it is not as clear as the secondary electrons, but the same adjustment is possible with the transmission electron image.

  Next, a procedure (S04 in FIG. 1, S04) of acquiring an image at each focus position by sequentially changing the focus within a range that sufficiently includes the sample surface and the lower surface will be described with reference to FIG. FIG. 5 shows the positional relationship between the sample cut out from the same direction as in FIG. 2A and the electron beam 308.

  That is, the sample is controlled by controlling the objective lens 304 by the focus adjustment position control unit 300c of the control device 300 using the electron beam condition set in S03 (focusing on the surface of the pattern by the secondary electron image). The transmission electron image (bright field image) 307 is obtained by changing the focus at a constant interval from slightly above the pattern surface (FIG. 1, S04). FIG. 5 shows an example of the beam position when the sample surface z0 is in focus and the focus position at which image acquisition is performed. The focus change interval p (15) may be set according to the interval of the cross section to be measured. As shown in FIG. 5, when the pattern length is L and the focus change interval is p, (L / p) +1 images are captured in the target pattern. At the time of actual image acquisition, the image is acquired by changing the focus so that it is sufficiently included below the sample length recorded at the time of sample creation. The focus control by the focus adjustment position control unit 300c can be easily performed by examining in advance the relationship between the control current amount of the objective lens 304 and the actual focus change amount. In general, the relationship between the control current amount of the objective lens 304 and the actual focus change amount can be obtained by design information of the STEM device or simulation based thereon and stored in the storage device 311, for example. Alternatively, a sample whose height is known in advance can be set in a holder, and the relationship between the control current amount of the objective lens 304 and the actual focus change amount can be examined in advance. Then, at each focus height, the deflector 303 is controlled by the electron beam scanning irradiation position control unit 300d to scan the electron beam 308 in the xy plane, and a transmission electron image is acquired by the bright field detector 307, for example. . Of course, a scattered electron image is acquired from the dark field detector 306. Although the y axis is not shown in FIG. 5, the depth direction on the paper surface is the y axis.

  Finally, the image processing unit 314 extracts the pattern edge position from the acquired transmission electron image under each focus condition by image processing, and calculates the pattern shape and dimension in the cross section of each focus position (FIG. 1, S05). ).

  FIG. 6 shows an example of an image acquired by the image processing unit 314 and a part of the image processing method. The figure shown in the upper part of FIG. 6 is a cut-out sample, and an example of a transmission electron image 17 obtained by enlarging a part of the transmission electron image (transmission electron image at a desired cross-sectional measurement position) 16 is shown on the lower side. FIG. As shown in FIG. 6, if the measurement target pattern is a dense line, it is possible to measure many lines at one time by preparing a single sample by capturing a plurality of transmission electron image images 16 by shifting the visual field. It becomes.

  Next, the image processing unit 314 detects the edge position of each pattern from the transmission electron image 17 shown in the center of FIG. 6 by image processing. FIG. 7 schematically shows the transmission electron signal waveform 18 of one line pattern in the scanning line of yi coordinate = ya in the image shown by the dotted line in FIG. In the transmitted electron image 17, a portion with a pattern becomes dark because a signal does not pass through, and a portion without a pattern becomes bright because an electron beam is detected as it is without being scattered. By detecting the boundary between the bright part and the dark part, the edge position of the pattern in the image is measured. Details of the edge extraction image processing in the STEM image will be described later. In the image of FIG. 6, first, the edge of the boundary with the pattern base (around yi = ys in FIG. 6) is detected. In this case, the signal waveform in the direction parallel to the yi axis is evaluated, and the edge position of the base is detected as in FIG. By using the detection result of the background boundary, if the image is rotated, rotation correction is performed on the image, or if the coordinate data of the measurement result is rotation corrected, a result without rotation is obtained. The edge coordinates of the pattern at each height with respect to the boundary with the ground are calculated by image processing. As for the edge of the upper surface of the pattern (around yi = yt in FIG. 6), the edge position is extracted by processing the waveform in the direction parallel to the yi axis as well as the boundary with the base.

  In the measurement of the present invention, there are a plurality of images similar to those in FIG. 6 but having different focus conditions. If all the pattern edge positions are calculated from these images, the three-dimensional structure of the target pattern can be obtained. FIG. 8 is an explanatory diagram of the three-dimensional structure information. As shown on the left side of FIG. 8A, in the measurement method of the present invention, a plurality of transmission electronic images 17 in which the focus condition (focus adjustment position) 14 is changed are obtained. By performing image processing on these, FIG. ) Edge position information 19 as shown on the right side can be obtained. In the diagram on the right side of FIG. 8A, each point on the outline has coordinate information of x and y. Since these edge coordinate information 20 indicates the value of the pattern outer peripheral edge at the coordinate z in the pattern longitudinal direction, the relative coordinate 14 of the focus position becomes the z coordinate between images. An embodiment displaying these three-dimensional structure data 21 is shown in FIG. It is preferable that the image and edge extraction results at each z coordinate shown in FIG. 8A and the three-dimensional stereoscopic display 21 as shown in FIG. 8B can be easily displayed on the display device 313. It is convenient that the viewing direction and magnification of the three-dimensional display can be easily changed by a mouse operation. Such display of three-dimensional data can be easily performed using a general CAD tool, a 3D animation tool, or the like.

  Next, a method for extracting pattern edges from a transmission electron image captured in various focus states will be described with reference to FIG. The inventors investigated how the amount of transmitted electron signals changes depending on the relationship between the scanning of the electron beam and the pattern edge position using simulation. FIG. 9 is a schematic diagram showing the relationship between the scanning of the electron beam 308, the edge position of the pattern 305, and the transmitted electron signal amount 18. 9A to 9C show the positional relationship between the electron beam and the sample sidewall, and the lower graph in FIG. 9 shows the relationship between each position and the signal amount of transmitted electrons. When a relatively thick sample is used, as shown in FIG. 9A, since the scattering of electrons starts from the place where the thickest part of the electron beam 308 contacts the pattern 305 for the first time, the signal amount starts to decrease. The amount of signal decreases gradually while the amount of electron beam hitting the target is small, but if the part where the beam density is focused and the beam crosses the sample, many electrons will be scattered, so the signal amount will drop sharply. (FIG. 9B). Thereafter, the electron beam 308 is all irradiated toward the inside of the sample 305, and reaches the minimum value when entering the inside until no scattered electrons are emitted from the side wall. Δx1 is determined by the position of (a), and becomes longer as the sample length L is thicker (longer) and the opening angle of the electron beam 308 is larger, and changes depending on the focus position. Further, as the sample length L is thicker (longer), the electron beam spreads in the sample due to scattering, so that the position (c) moves to the right and Δx2 increases. In order to always measure the height of the focused image in the image with the focus changed, it is only necessary to correctly detect the edge position corresponding to FIG. However, since the signal amount 18 varies greatly between the edges (a) to (c) of the pattern, and particularly between (a) and (b), the measurement error depends on which part is extracted as the pattern edge. May become large. According to the inventors' investigation, it is known that the distance between (a) and (b) can be a value exceeding 10 nanometers, although it depends on the opening angle of the electron beam 308 and the thickness L of the sample 305. Since the size of the semiconductor pattern to be evaluated by the present invention is about several tens of nanometers, and the dimensional variation to be taken into account due to roughness is several nanometers, it affects the signal degradation between (a) and (b). It is necessary to detect an edge position that is not affected. In addition, the signal amount sb in (b) varies depending on the focus position. As shown by the solid line 18a, the closer the focus position is to the sample surface, the larger the signal amount in (b). However, the closer the focus is to the lower surface (in the case indicated by the dotted line 18b), the more scattered the irradiated electrons. Therefore, the signal amount is reduced. The correct edge position cannot be determined without performing image processing in consideration of such a physical phenomenon.

  Next, a method for extracting a correct edge position from such an image in the image processing unit 314 will be described with reference to FIG. FIG. 10 shows the transmission signal waveform 18 and its first derivative waveform 22 near the left edge of the sample shown in FIG. Since the signal amount rapidly decreases in the portion corresponding to (b), if the portion 23 where the differential waveform 22 has the minimum value (maximum value in the case of the right edge) is detected as an edge, the slope is close to (b) Can detect the maximum point. Therefore, in the present invention, the coordinates that take the minimum value (in the case of the left edge) and the maximum value (in the case of the right edge) of the first derivative are used as the edge position, so that the pattern edge can be stably and accurately irrespective of the focus change. Can be obtained. Similarly, by detecting that the third derivative zero cross point immediately before the first derivative peak using the fact that the third derivative becomes zero at the inflection point, the point where the signal amount rapidly decreases is detected. Is also possible.

  Next, another embodiment of the edge position extraction method will be described with reference to FIG. In FIG. 10, the edge position is determined using the peak of the first derivative, but the differential waveform has a drawback that it is vulnerable to noise. In order to acquire an image with good noise, it takes a lot of time to acquire the image. Moreover, since the number of electrons to be irradiated increases, problems such as damage to the sample may occur. Therefore, in the method of FIG. 11, a threshold value corresponding to the focal position is set by utilizing the fact that the signal amount in (b) decreases as the focal position becomes lower, and the intersection of the set threshold value and the waveform. Is the edge position. In the embodiment of FIG. 11, for example, when the focal position is relatively high and a waveform 18a as drawn with a solid line is obtained, the amplitude A of the transmitted electron image (the darkest part of the background portion and the pattern portion). For example, when the intersection point E1 with the threshold value of 60% is the edge position 26a, the focal point is at the lower side, and a waveform 18b like a dotted line is obtained, the threshold value Is set to 30%, for example, and the intersection point E2 with the waveform is set as the edge position 26b. It is necessary to set appropriate values for these threshold values 25a and 25b according to the focal position. This can be done by acquiring a waveform in advance using a sample whose shape is known. The method of FIG. 11 needs to determine the threshold value in advance, but has the advantage of being more resistant to noise than the method of FIG.

  Next, a method for determining an appropriate beam opening angle for executing the pattern measurement method of the present invention will be described with reference to FIG. As shown in FIG. 9, in the pattern measurement method of the present invention, the signal amount sharply decreases at the intersection of the sample 305 and the focus point of the electron beam 308. However, if the electron beam 308 irradiated above the focus point crosses the sample 305, the electrons are scattered in the sample and the actual resolution is lowered. According to the inventors' investigation, it is known that the measurement accuracy is improved when a part of the electron beam is directly irradiated on the sample at the focus point. Therefore, an appropriate beam opening angle α for a sample having roughness may be determined based on the roughness of the line and set for the beam opening angle adjusting unit 300a. In order to determine an appropriate opening angle, the line pattern is observed from above (a state where the upper surface of the pattern can be seen as shown in FIG. 2A). The observation may be performed using the SEM at the time of FIB processing, or is measured in advance by a length measurement SEM, AFM, or the like, and is input via, for example, the I / F 312 and stored in the storage device 311. Next, the image processing unit 314 detects an edge position by observing the line stored in the storage device 311 from above. The edge position detection here needs only to know the relative position. When detecting the edge position, the image processing unit 314 obtains, for each point on the edge, a straight line that passes through the point and circumscribes the line outside the line. For example, two straight lines 29 are obtained for the point 28 shown in FIG. Further, the image processing unit 314 obtains the inclinations θ1 and θ2 when the longitudinal direction 27 of the line is used as a reference, and records the larger one in the storage device 311, for example. Further, the image processing unit 314 performs this for a line edge having a certain length, and estimates the maximum angle θmax. The beam opening angle α (30) is set to a value larger than θmax estimated by the image processing unit 314 by controlling the multi-stage electron lens 302 and the like based on a command from the beam opening angle adjusting unit 300a of the control device 300. By doing so, even if the sample has roughness, the beam can reach the sample 11 (305) at the focus position while maintaining a certain density without being scattered within the solid. For example, when electrons are incident in the direction indicated by the arrow in FIG. 12, the beam reaches the surface of the sample 11 at the focus position without being scattered by the pattern.

  As described above, the STEM image processing unit 314 can perform many cross-sectional shape measurements in a relatively short time by the measurement method according to the first embodiment described with reference to FIGS. The pattern shape information 20 as shown in FIG. These results are useful as pattern three-dimensional information as they are, and based on this result, the image processing unit 314 has the line pattern longitudinal direction (z direction shown in FIG. 2B) and height direction (FIG. 2). It is possible to analyze in detail how the dimensions fluctuate with respect to the y direction (2 (b)). For example, you can check how the line edge roughness (LER (Line Edge Roughness)) and line width roughness (LWR (Line Width Roughness)) change for each line pattern height (z value) It can be used to check the change of the tilt angle. That is, according to the present invention, with respect to a measurement target pattern consisting of a line pattern having a length of 2 μm or more in the z direction, the average line pattern dimension, line width roughness, line edge roughness in a predetermined cross section (predetermined z value). Can be measured.

  As described above, for the semiconductor circuit pattern to be measured, the image processing unit 314 can evaluate the manufacturing variation of the cross-sectional shape in the z direction, which has been difficult until now. In particular, even for a shape change such as a reverse taper or a notch that has been difficult to grasp with a nondestructive measurement method, a highly accurate shape measurement can be performed by using a STEM image with a high resolution. These results can be used as a management tool for semiconductor manufacturing processes and as an analysis tool for investigating the cause of abnormalities in the process. Quantitative evaluation can be performed in a short period of time, and the start-up period can be shortened. Further, since it is possible to perform statistical evaluation in consideration of variations in cross-sectional shape and dimensions, it is possible to apply to calibration of various measuring devices used in semiconductor lines.

[Second Embodiment]
In the first embodiment according to the present invention, the case where there is no protective film around the measurement target pattern has been described. In the second embodiment of the present invention, a case where a protective film is provided around the measurement target pattern will be described. As described in the embodiment of FIG. 12, in order to perform the pattern measurement of the present invention with high accuracy, it is desirable that as many electrons as possible converge on the surface of the measurement object without being scattered by the solid. . However, depending on the material, there is a case where it is difficult to process the cross section without a protective film. In such a case, it is necessary to reduce the decrease in accuracy due to the influence of the protective film. FIG. 13 shows an example of how the convergent beam expands due to scattering within the solid. When a 200 keV electron beam is irradiated, the left figure shows the beam spreads 31 and 32 when focused to a depth of 2 microns from the sample surface and the right figure is 1 micron. When there is no scattering in the solid, the calculation is performed under the condition that the beam diameter is infinitely small. As can be seen from FIG. 13, the effective diameter of the electron beam increases as the depth from the sample surface increases. The figure disclosed in Fig. 3 of Patent Document 1 is similar data, but it is not proportional to the depth, but the resolution suddenly drops to some extent. This effect is unavoidable when there is a protective film. In the second embodiment according to the present invention, a method corresponding to the increase in the beam diameter due to the diffusion in the solid is disclosed.

  FIG. 14 illustrates measurement variations when a thick sample is measured. When measuring a thick sample, the measurement dispersion increases as the distance from the sample surface increases due to the diffusion of electrons in the solid and the decrease in signal amount and S / N at the edge position. When the same sample is repeatedly measured, as shown in FIG. 14, the variation in the detected edge position 26 increases as it goes down. Therefore, in the STEM, as shown in FIG. 15, once the sample 11 (305) is measured, the sample 11 (305) is rotated 180 degrees and the measurement is performed again in the same manner. In FIG. 15A, the edge position on the A side can be accurately measured in the first measurement, and the edge position on the B side can be accurately measured in the second measurement. Therefore, the surface as shown in FIG. By performing a weighted average for the detected edge position using the S-shaped weight function 33 according to the distance from the distance, it is possible to measure a stable cross-sectional shape regardless of the pattern thickness (length L). A generally used sigmoid curve may be used for the S-shaped curve. FIG. 16 shows an evaluation procedure in the second embodiment of the STEM. A sample is prepared in the same manner as in the first embodiment shown in FIG. 1, and the sample is set in a STEM holder (S01, S02). Next, after adjusting the surface of the pattern by focusing with the secondary electron image (S03), an image set with the focus changed is acquired (S04). In the case where the entire image is captured and the double-sided image is not acquired (S06), the sample is rotated 180 degrees (S07), and the image processing unit 314 similarly acquires the image. After detecting the edge position in each image (S05), the image processing unit 314 combines the positions of the two measurement results (S08). Further, the image processing unit 314 will execute what will be described next. That is, first, the coordinates in the y direction are matched based on the background position. Next, roughly align the positions based on the focus information combined in the secondary electron image, and match the roughness using the pattern measurement results in the center in the z direction where the accuracy of the two measurement results is the same. What is necessary is just to match the position of x and z. For the fitting, the shift may be performed so that the sum of squares of differences between a plurality of edge coordinates in a plurality of images is minimized. Finally, weighting is performed according to the distance from the surface as shown in FIG. 15B, and the pattern size is calculated by weighted averaging (S09). In addition, when the thickness which can be measured is not enough, it can respond by raising the acceleration voltage in the acceleration voltage setting part 300b. As described above, the second embodiment according to the present invention is different from the first embodiment in that steps S06, S07, S08, and S09 are provided.

  Of course, this method may be applied to a sample without a protective film. As shown in FIG. 9, the lower the focus position, the lower the signal amount at the edge portion and the worse the S / N for edge extraction. Therefore, by performing measurement by combining the results measured from above and below by the method shown in FIGS. 15 and 16, it is possible to reduce measurement value variation due to S / N reduction and realize more accurate measurement.

[Third Embodiment]
The difference of the third embodiment according to the present invention from the first and second embodiments is that, for example, the image processing unit 314 performs tilt correction. Next, a description will be given of a method of correcting the inclination of the sample generated when the measurement target pattern is attached, which is the third embodiment in pattern shape measurement by the STEM according to the present invention, and outputting the corrected result. At the time of observation and measurement with the STEM, the sample processed with the FIB is observed with the STEM. However, depending on how this sample is attached, the longitudinal direction of the line may not be parallel to the z-axis. (Refer FIG.2 (b)). In such a case, the measurement result has an inclination as shown in FIG. When the image processing unit 314 has such an inclination, the image processing unit 314 obtains a linear line 34 that approximates the data (detected edge position) 26 by the least square method, and corrects the measurement data. Just do it. If the tilt can be canceled by correcting the data in this way, a dimension measurement error caused by the tilt can be eliminated, and more accurate measurement can be performed. Further, as shown in FIG. 8B, there is an advantage that it is easy to handle when the pattern structure 21 is represented in three dimensions. In addition, when the observation direction is changed as in the second embodiment, the inclination with respect to the z-axis may change before and after the rotation depending on the accuracy of the stage. After correcting the above, processing such as a weighted average may be performed.

  However, if there is low frequency roughness, the estimation of the slope may have a large error. Therefore, as shown in FIG. 17B, it is better to use more data using other measurement points, for example, left and right edge roughness and data of different heights (y coordinates) as necessary. The estimation accuracy of the slope is increased. Since data with only different heights often has a correlation in roughness, it is desirable to include information on left and right or adjacent lines. Although FIG. 17 shows the inclination in the x direction, it goes without saying that the same can be done in the y direction.

  It is generally very difficult to hold the sample 305 processed by FIB on a STEM holder (not shown) without tilting. However, as shown in the third embodiment of the present invention, if correction is performed, highly accurate measurement can be realized even if the sample is held tilted.

[Fourth Embodiment]
The difference of the fourth embodiment according to the present invention from the first to third embodiments is that the image processing unit 314 performs processing for removing abnormal values.

  Next, a method for further improving measurement accuracy will be described with reference to FIG. 18 except for an abnormal value which is the fourth embodiment in pattern shape measurement by the STEM according to the present invention. In the pattern shape measuring method of the present invention, the focal position is sequentially changed in the z direction, and the xy cross-sectional shape is measured at each focal position. Since the cross-sectional shape in the z direction of the normal measurement target pattern changes continuously, the measurement stability can be improved by utilizing the continuity. In FIG. 18, white circles indicate the outlier 35 of the xy measurement. When a semiconductor line pattern is to be measured, since the roughness of the line pattern is usually evaluated by a length measurement SEM or CD (Critical Dimension) SEM (not shown), the length measurement SEM or CDSEM (not shown) is used. For example, the image processing unit 314 often knows the approximate range of fluctuation because the data is input via the I / F 312 and stored in the storage device 311. Therefore, the image processing unit 314 can remove, as an outlier, a value that deviates more than the variation of the line itself from the fluctuation in the obtained cross-sectional measurement result (detected edge position) in the z direction. In the embodiment of FIG. 18, an average value 36 in the z direction of the edge position 26 on one side is calculated, and the tolerance corresponding to the value of the line edge roughness that has been separately evaluated in a length measurement SEM or CDSEM (not shown). A range 37 is set, and values that are out of the range 37 are removed from the measurement result. In addition to the embodiment of FIG. 18, it is also possible to perform a process of calculating a moving average of the three-dimensional measurement results of several points of the line pattern and removing those that are greatly deviated therefrom. After removing the outlier, the original edge position may be estimated from the previous and subsequent measurement results (in any direction of x, y, z).

  As described above, according to the fourth embodiment of the present invention, a transmission electron image or scattering obtained by imaging a measurement target pattern formed as a line pattern in the z direction by sequentially changing the focal position. In the three-dimensional cross-sectional shape measurement result measured at the edge position 26 at each focal position based on the electronic image, it is further stable by removing an outlier that is out of the allowable range 37 corresponding to the value of the line edge roughness. Three-dimensional cross-sectional shape measurement is possible.

[Fifth Embodiment]
A difference of the fifth embodiment according to the present invention from the first to fourth embodiments is a GUI in the image processing unit 314.

  Next, an example of a GUI for inputting a condition for image acquisition and processing, which is a fifth embodiment of pattern shape measurement by STEM according to the present invention, and a result display will be described. FIG. 19 shows an embodiment of the input condition (condition input screen) 315. In the three-dimensional cross-sectional shape measuring method according to the present invention, the basics such as the acceleration voltage (keV), the magnification (kx) of the image to be acquired, and the depth of focus mode (electron beam opening angle α) according to the desired resolution in the z direction are used. Image acquisition conditions, types of images to be used (transmission electron images, scattered electron images, reflection electron images (secondary electron images, etc.)), and the range of image acquisition with a focus ((z0-3) to ( z0 + (L / p) +2)) for determining the pattern length L (μm), measurement pitch p (nm), and image processing conditions (filter processing conditions and differential processing conditions shown in FIG. 10) used for edge position extraction Alternatively, the threshold value processing condition shown in FIG. Here, it is necessary to put a value L, which is measured in advance, smaller than the dimension of the measurement target sample, with respect to the range in which image acquisition is performed. On the other hand, FIG. 20 shows an example of a measurement result display (result display screen) 316. As shown on the left side of FIG. 20, a three-dimensional display (three-dimensional graph) that can be operated interactively, which is a three-dimensional cross-sectional shape measurement result, is displayed so that the measured point can be selected by moving the mouse closer. When a measurement point is clicked, an image (such as a reflected electron image 38 and a transmitted electron image 17) obtained by interactively calculating data on the location is displayed on the right side. In the embodiment of FIG. 20, the representative signal waveforms (the reflected electron signal waveform 39 and the transmitted electron signal waveform 18) and the measurement results are displayed on the image.

  Thus, by displaying the GUI screens 315 and 316 as shown in FIG. 19 and FIG. 20 on the display device 113, it becomes possible to easily perform condition setting and result confirmation.

[Sixth Embodiment]
The sixth embodiment according to the present invention is different from the first to fifth embodiments in that, for example, in the image processing unit 314 of the STEM, other measurement tools (measurement SEM or CDSEM or AFM for measurement) are used. ) And calibrate.

  Next, a method for calibrating with another measurement tool according to the sixth embodiment in the pattern shape measurement by the STEM according to the present invention will be described. In conventional cross-sectional measurement, only one cross-sectional dimension can be measured at a time, and it has been difficult to obtain a large amount of measurement data. However, if the measurement method of the present invention is used, a large amount of data can be obtained by creating a sample once. Therefore, even when there is a variation in the shape of the measurement object, other measurement devices (for length measurement SEM or CDSEM or for length measurement) AFM) and the measurement result can be compared for calibration.

  FIG. 21 shows an example in which the measurement results are compared with the length measurement SEM or CDSEM. When calibrating, it is necessary to measure and compare the same sample using different methods. Therefore, first, taking into account the accuracy of cutting out the cross-sectional sample, measurement is performed by a length measuring SEM or CDSEM as described in JP-A-2005-265424. As shown in FIG. 21, when there is the line pattern 11 and the alignment pattern 40, the FIB processing accuracy 41 that can be cut out using the alignment pattern 40 is checked in advance. Then, all the patterns that may be displaced by this FIB cutout are measured by the length measurement SEM or CDSEM (area 42). Next, cut-out processing is performed by FIB, and for example, cross-sectional shape measurement of 43 regions is performed by the method described in the first to sixth embodiments. The measurement result of the area 43 is compared with the length measurement SEM measurement result in the area 42 measured in advance, and the matching area is determined. Here, the alignment is performed using a relative change in the edge position due to the roughness instead of the dimension. In this way, if calibration between devices is performed using the measurement result of the length measurement SEM or CDSEM of the portion that coincides with the cross-sectional measurement position by the STEM, it is possible to perform highly accurate calibration in consideration of pattern variations. Become. Even when the pitch in the z direction measured in the z-x plane by the length measuring SEM or CDSEM and the pitch in the z direction that can be measured by the cross-sectional measurement by the STEM of the present invention are different, both are random in the same population. Since it can be considered as sampling, there is an advantage that statistical treatment becomes easy. In the embodiment of FIG. 21, the case of the length measurement SEM or the CDSEM has been described, but the same processing is possible for a length measurement AFM (Atomic Force Microscope) or the like.

  As described above, since the cross-sectional shape measurement method according to the present invention can be used for statistical evaluation in consideration of variations in cross-sectional shape and dimensions, it can be used for calibration of various measuring devices used in semiconductor lines. It becomes possible to apply.

  In the embodiments so far, a pattern cross-sectional shape measuring method using a general STEM has been disclosed. As a system that can further improve the accuracy, the aberration corrector disclosed in Non-Patent Document 5 is introduced into the STEM to realize a beam with a shallow focal depth and a high electron density at the in-focus position. be able to. If such an electron beam can be realized, the resolution of the STEM image is increased and the measurement accuracy is improved. In addition, since the depth of focus is shallow, the resolution in the z direction is improved, and cross-sectional measurement at finer intervals becomes possible.

  As another system capable of improving the measurement accuracy, FIB using light ions disclosed in Non-Patent Document 6 may be used instead of the electron beam. When FIB is used, since the scattering angle in a solid is relatively small, measurement with good resolution becomes possible.

  In the present invention, the measurement method using transmission electrons has been described. However, the same shape measurement can be performed using scattered electrons. In addition, measurement can be performed with higher accuracy by measuring each with transmitted electrons and scattered electrons and integrating the results.

  The pattern measurement technique of the present invention can be applied in the same manner as long as it can acquire an image with a transmission electron microscope and has a shape with a relatively small height difference such as a rectangle. In the present invention, the measurement of the semiconductor pattern has been described, but the present invention can be applied to MEMS, fine industrial parts, and the like.

It is a figure for demonstrating the procedure of 1st Embodiment of the pattern measurement method using STEM which concerns on this invention. It is a figure for demonstrating the measurement object pattern structure and coordinate definition which concern on this invention. It is a schematic diagram of a measurement target line pattern when there is a line edge roughness according to the present invention. It is a system explanatory drawing of STEM which implements the pattern measurement method concerning this invention. It is a figure for demonstrating the focus position setting method of the pattern measurement method which concerns on this invention. It is a figure which shows the Example of the image acquisition of the pattern measurement method which concerns on this invention. It is explanatory drawing of the transmission electron image and edge extraction which concern on this invention. It is explanatory drawing of the image acquired based on this invention, and a processing result. It is a figure explaining the characteristic of the transmission electron image used with the pattern measurement method which concerns on this invention. It is explanatory drawing of the edge extraction image process in the pattern measurement method which concerns on this invention. It is explanatory drawing of the edge extraction image process in the pattern measurement method which concerns on this invention. It is explanatory drawing of the method of setting the opening angle of an electron beam from the measurement object pattern shape which concerns on this invention. It is explanatory drawing which shows the relationship between the pattern thickness which concerns on this invention, and an effective beam diameter. It is a figure explaining the change of the measurement dispersion | variation by the pattern thickness which concerns on this invention. It is explanatory drawing of the method of measuring the thick pattern based on this invention accurately. It is explanatory drawing which shows the procedure of the method of measuring the thick pattern based on this invention accurately. It is a figure for demonstrating the method to correct | amend the inclination of the pattern which concerns on this invention. It is a figure for demonstrating the method of removing the outlier which concerns on this invention. It is explanatory drawing of GUI which sets the image acquisition and measurement condition which concern on this invention. It is explanatory drawing of GUI which displays the measurement result which concerns on this invention. It is a sampling method explanatory drawing in the case of calibration with the other measuring device which concerns on this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 ... Line pattern, 11a ... Roughness line pattern, 12, 12a ... Base, 13 ... Sample cut-out area, 14 ... Focus adjustment position, 15 ... Measurement interval p, 16 ... Transmission electron imaging area, 17 ... Transmission electron image, 18 ... Transmission electron signal waveform, 19 ... Edge extraction result, 20 ... Edge coordinate information, 21 ... Measurement result three-dimensional display, 22 ... Transmission electron signal primary differential waveform, 23 ... Minimum value peak, 24 ... Signal waveform amplitude A, 25 ... Threshold value 26 ... detected edge position, 27 ... line longitudinal direction (reference), 28 ... edge position (Top view, AFM measurement result, etc.), 29 ... roughness tangent, 30 ... beam opening angle α, 31 ... electron distribution at the in-focus height, 32 ... electron beam intensity distribution at the in-focus height, 33 ... weight function (sigmoid function), 34 ... approximate straight line by least square method, 35 ... off 36: average value, 37: allowable range, 38: reflected electron image, 39: reflected electron signal waveform, 40: alignment mark, 41: cross-sectional measurement sample cutout position accuracy, 42: measurement SEM (or AFM) measurement range , 43 ... Section measurement area,
DESCRIPTION OF SYMBOLS 300 ... Control apparatus, 300a ... Beam opening angle adjustment part, 300b ... Acceleration voltage setting part, 300c ... Focus adjustment position control part, 300d ... Beam scanning irradiation position control part, 301 ... Electron source, 302 ... Electron lens, 303 ... Deflection 304, objective lens, 305 ... specimen (sample), 306 ... dark field detector, 307 ... bright field detector, 308 ... electron beam, 309 ... A / D converter, 310 ... secondary electron detector, 311 ... storage device, 312 ... external interface, 313 ... display device, 314 ... image processing unit, 315 ... condition input screen (GUI), 316 ... result display screen (GUI).

Claims (25)

A pattern measurement method for measuring a three-dimensional cross-sectional shape of a measurement target pattern using a scanning transmission charged particle microscope,
The scanning irradiation position of the focused charged particle beam is aligned with the desired xy cross-section measurement position of the measurement target pattern, and the focal position of the focused charged particle beam is sequentially changed with respect to the z direction of the measurement target pattern. An electronic image acquisition process for acquiring a transmission electron image or a scattered electron image of the measurement target pattern at a focal position;
In the electronic image acquisition process, a transmission electron image or a scattered electron image acquired at each focal position matched with the desired xy cross-sectional measurement position is processed, and measurement in the electronic image at the respective focal position is performed. An edge position calculation process for calculating the edge position of the target pattern;
The measurement based on a combination of the edge position of the measurement target pattern in the electronic image at each focal position calculated in the edge position calculation process and the focal position from which the electronic image was acquired in the electronic image acquisition process. A pattern measuring method comprising: a cross-sectional shape measuring process for measuring a three-dimensional cross-sectional shape of a target pattern.   The pattern measurement method according to claim 1, wherein the measurement target pattern is formed as a line pattern in the z direction. The measurement target pattern is formed as a line pattern having a length of 2 μm or more in the z direction,
2. The cross-sectional shape measurement process, wherein the average dimension, line width roughness, or line edge roughness of the line pattern in a predetermined z-section is measured as a three-dimensional cross-sectional shape measurement of the measurement target pattern. Pattern measurement method.   The pattern measurement method according to claim 1, wherein a solid such as a protective film is not attached around the measurement target pattern.   The pattern measurement method according to claim 1, wherein, in the electronic image acquisition process, an acceleration voltage of the charged particle beam in the scanning charged particle beam microscope is 200 keV or more.   The electronic image acquisition step further includes a beam opening angle determination step of determining a beam opening angle of the charged particle beam in the scanning charged particle beam microscope based on a dimensional change amount of the measurement target pattern. The pattern measurement method according to claim 1.   2. In the edge position calculation process, in a signal waveform obtained from an electronic image at each of the focal positions, a coordinate having the maximum inclination is set as an edge coordinate of the measurement target pattern. The pattern measuring method described in 1.   In the edge position calculation process, in the signal waveform obtained from the electronic image at each of the focal positions, the point where the signal waveform matches the threshold set according to the image acquisition condition is the edge of the measurement target pattern. The pattern measuring method according to claim 1, wherein coordinates are used.   In the cross-sectional shape measurement process, the relationship between the calculation result of the edge position of the measurement target pattern in the electronic image at each focus position calculated in the edge position calculation process and the focus position is approximated by a linear expression. The pattern measurement method according to claim 1, further comprising an inclination correction step of correcting a calculation result of an edge position of the measurement target pattern at each focus position based on the inclination of the approximate expression.   In the cross-sectional shape measurement process, the difference from the average value of the calculation result of the edge position of the measurement target pattern in the electronic image at each focal position calculated in the edge position calculation process is larger than a predetermined tolerance value. The pattern measurement method according to claim 1, further comprising: an exclusion process that excludes a three-dimensional cross-sectional shape measurement result of the measurement target pattern when the measurement target pattern is large. A pattern measurement method for measuring a three-dimensional cross-sectional shape of a measurement target pattern using a scanning charged particle beam microscope,
The scanning irradiation position of the focused charged particle beam is matched with the desired xy cross-section measurement position of the measurement target pattern, and the focal position of the focused charged particle beam is sequentially changed in the z direction with respect to the measurement target pattern. A first electronic image acquisition process for acquiring a first transmission electron image or a scattered electron image of the measurement target pattern at a focal position;
In the first electronic image acquisition process, the first transmission electron image or the scattered electron image acquired at each focal position adjusted to the desired xy cross-sectional measurement position is processed, and at the respective focal position, A first edge position calculation process for calculating the edge position of the measurement target pattern in the electronic image;
The measurement target pattern is obtained by inverting the measurement target pattern and inverting the focal position of the converged charged particle beam by aligning the scanning irradiation position of the focused charged particle beam with the desired xy cross-section measurement position of the inverted measurement target pattern. A second electronic image acquisition process in which the second transmission electron image or the scattered electron image of the inverted measurement target pattern at each focal position is acquired sequentially in the z direction,
In the second electronic image acquisition process, the second transmission electron image or the scattered electron image acquired at each focal position adjusted to the desired xy cross-sectional measurement position is processed, and A second edge position calculation process for calculating the edge position of the inverted measurement target pattern in the electronic image;
The combination of the edge position of the measurement target pattern in the electronic image at each focal position calculated in the first edge position calculation process and the respective focal position from which the electronic image was acquired in the first electronic image acquisition process And the edge position of the inverted measurement target pattern in the electronic image at the respective focal positions calculated in the second edge position calculating process and the respective focal positions from which the electronic image was acquired in the second electronic image acquiring process And a cross-sectional shape measurement process for measuring a three-dimensional cross-sectional shape of the measurement target pattern based on the combination with the pattern measurement method.   In the cross-sectional shape measurement process, the calculation result of the edge position of the measurement target pattern calculated in the first edge position calculation process and the edge position of the inverted measurement target pattern calculated in the second edge position calculation process The pattern measurement method according to claim 11, wherein the three-dimensional cross-sectional shape measurement of the measurement target pattern is performed by a weighted average obtained by applying a weight corresponding to the depth from the surface of the measurement target pattern to the calculation result. .   In the cross-sectional shape measurement process, the relationship between the calculation result of the edge position of the measurement target pattern in the electronic image at each focal position calculated in the first edge position calculation process and the respective focus positions, and the first Each of the relationship between the calculation result of the edge position of the inverted measurement target pattern in the electronic image of each focus position calculated in the edge position calculation process of 2 and the respective focus positions is approximated by a linear expression, 12. The method according to claim 11, further comprising an inclination correction step of correcting each of the calculation result of the edge position of the measurement target pattern and the calculation result of the edge position of the inverted measurement target pattern by the inclination of each approximate expression. The pattern measurement method described.   In the cross-sectional shape measurement process, the difference from the average value of the calculation result of the edge position of the measurement target pattern in the electronic image at each focal position calculated in the first edge position calculation process is a predetermined tolerance. The difference from the average value of the calculation result of the edge position of the inverted measurement target pattern in the electronic image of each focal position calculated in the second edge position calculation process when the value is larger than the value is designated in advance. The pattern measurement method according to claim 1, further comprising: an exclusion process that excludes a three-dimensional cross-sectional shape measurement result of the measurement target pattern when the value is larger than an allowable value.   12. The display method according to claim 1, further comprising: a display step of displaying a three-dimensional cross-sectional shape measurement result of the measurement target pattern measured in the cross-sectional shape measurement step as a three-dimensional graph that can be operated interactively. The pattern measurement method described. A scanning transmission charged particle microscope for measuring a three-dimensional cross-sectional shape of a measurement target pattern,
The focus position of the focused charged particle beam is sequentially changed with respect to the z direction of the measurement target pattern by matching the scanning irradiation position of the focused charged particle beam with the desired xy cross-section measurement position of the measurement target pattern, and the respective focal points. An electronic image acquisition unit for acquiring a transmission electron image or a scattered electron image of the measurement target pattern at a position;
The electronic image acquisition unit processes a transmission electron image or a scattered electron image acquired at each focal position that matches the desired xy cross-section measurement position, and measures the electronic image at the respective focal position. An edge position calculation unit that calculates the edge position of the target pattern, and the edge position of the measurement target pattern in the electronic image at each focal position calculated by the edge position calculation unit and the electronic image acquired by the electronic image acquisition unit A scanning transmission charged particle microscope comprising: an image processing unit having a cross-sectional shape measuring unit that measures a three-dimensional cross-sectional shape of the measurement target pattern based on a combination with each focal position.   The cross-sectional shape measurement unit of the image processing unit is configured to measure an average dimension, line width roughness, or line edge roughness of a line pattern in a predetermined z-section as a three-dimensional cross-sectional shape measurement of the measurement target pattern. The scanning transmission charged particle microscope according to claim 16.   The scanning according to claim 16, further comprising a beam opening angle determining unit that determines a beam opening angle of the charged particle beam based on a dimensional change amount of the measurement target pattern in the electronic image acquisition unit. Type transmission charged particle microscope.   The edge position calculation unit of the image processing unit is configured such that a coordinate having the maximum inclination is used as an edge coordinate of the measurement target pattern in a signal waveform obtained from the electronic image at each focal position. The scanning transmission charged particle microscope according to claim 16.   In the edge position calculation unit of the image processing unit, in the signal waveform obtained from the electronic image at the respective focal positions, the point at which the signal waveform matches the threshold set according to the image acquisition condition is measured. The scanning transmission charged particle microscope according to claim 16, wherein the scanning transmission charged particle microscope is configured to be edge coordinates of a target pattern.   In the cross-sectional shape measurement unit of the image processing unit, the relationship between the calculation result of the edge position of the measurement target pattern in the electronic image at each focal position calculated by the edge position calculation unit and each focal position is 1 The scanning transmission according to claim 16, further comprising an inclination correction unit that approximates the following expression and corrects the calculation result of the edge position of the measurement target pattern at each focal position by the inclination of the approximation expression. Charged particle microscope. A scanning transmission charged particle microscope for measuring a three-dimensional cross-sectional shape of a measurement target pattern,
The scanning irradiation position of the focused charged particle beam is matched with the desired xy cross-section measurement position of the measurement target pattern, and the focal position of the focused charged particle beam is sequentially changed in the z direction with respect to the measurement target pattern. A first transmission electron image or a scattered electron image of the measurement target pattern at a focal position is acquired, and the measurement target pattern is reversed, and the scanning irradiation position of the converged charged particle beam is reversed. The focus position of the focused charged particle beam is sequentially changed in the z direction with respect to the inverted measurement target pattern in accordance with the xy cross-section measurement position of the second, and the second of the inverted measurement target pattern at each focus position. An electronic image acquisition unit for acquiring a transmission electron image or a scattered electron image of
In the electronic image acquisition unit, the first transmission electron image or the scattered electron image acquired at each focal position matched with the desired xy cross-sectional measurement position is processed, and the electronic image at the respective focal position is processed. The edge position of the measurement target pattern is calculated, and further, the second transmission electron image or the scattered electron image acquired at the respective focal positions matched with the desired xy cross-section measurement position in the electronic image acquisition unit is processed. Then, an edge position calculation unit that calculates an edge position of the inverted measurement target pattern in the electronic image at each focal position, and an electronic image at each focal position calculated by the edge position calculation unit The combination of the edge position of the measurement target pattern and the respective focal positions, and the reflection in the electronic image at each focal position calculated by the edge position calculation unit. And an image processing unit including a cross-sectional shape measuring unit that measures a three-dimensional cross-sectional shape of the measurement target pattern based on a combination of the edge position of the measurement target pattern and the respective focus positions. Scanning transmission charged particle microscope.   In the cross-sectional shape measurement unit of the image processing unit, the calculation result of the edge position of the measurement target pattern calculated by the edge position calculation unit and the calculation result of the edge position of the inverted measurement target pattern calculated by the edge position calculation unit 23. The scanning according to claim 22, wherein three-dimensional cross-sectional shape measurement of the measurement target pattern is performed by a weighted average obtained by applying a weight in accordance with a depth from the surface of the measurement target pattern to each other. Type transmission charged particle microscope.   Further, the electronic image acquisition unit includes an image acquisition condition including a longitudinal dimension of the measurement target pattern for acquiring a transmission electron image or a scattered electron image of the measurement target pattern and an interval between the focus positions to be sequentially changed, and The scanning transmission charged particle microscope according to claim 16 or 22, further comprising a graphical user interface for inputting and setting an image processing condition for calculating an edge position of a measurement target pattern in the image processing unit. . A first step of obtaining a first measurement result by performing measurement in a wider range than the variation of the cross-sectional sample extraction position in advance by another measurement method to be compared;
A second step of obtaining a second measurement result by performing a three-dimensional cross-sectional shape measurement by the pattern measurement method according to any one of claims 1 to 11;
Of the first measurement results in the first step, a third step of comparing the measurement results using only the measurement results in the region that matches the second measurement result in the second step A calibration method between measuring apparatuses, comprising:

Images